Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Discovery Grants Program - Individual

University of New Brunswick

$31,000.00

New Efficient Cryptographic Methods and Protocols for Privacy-preserving Machine Learning

EF 2024-25

Résumé
Cybersecurity and privacy is a growing concern in our society that includes business, people and governments. In the present age of the Fourth Industrial Revolution (Industry 4.0), extracting knowledge or value from enormous amounts of data is used to make smart decisions in business intelligence. Machine learning (ML) can learn deeper knowledge/insight from data, and in particular, deep learning has the huge potential to learn from large datasets. As more and more data makes ML powerful, unlocking values from data using ML requires sharing, processing and analyzing individual's (personal) data between organizations, which puts the individuals' rights to privacy at risk. In recent years, industries have been developing ML systems (e.g., MLaaS) for various tasks such as image classification, predictive analysis, and language models (e.g., ChatGPT). To build trustworthy and strongly secure ML systems and accelerate their deployment, security and privacy issues must be addressed, while considering various threats from adversarial machine learning attacks and cyberattacks to the ML systems. Regulations such as HIPPA and GDPR have been in place to control sharing sensitive user information, and organizations are coerced to embrace privacy enhancing technologies to protect data and users' privacy. In this research program, I propose to develop new, efficient, provably secure cryptographic techniques and security protocols to preserve the confidentiality and integrity in ML computations and develop diverse countermeasures against different cybersecurity and specialized adversarial ML attacks in an ML system. Our first objective is to design communication-efficient and secure ML training and verifiable cryptographic inference protocols to provide data and model privacy for classification and regressor models such as linear, neural networks and language models. Our second objective is to investigate new constructions of efficient hybrid homomorphic encryption schemes and multi-party threshold cryptography schemes for lightweight ciphers as well as standardized symmetric-key ciphers and develop new end-to-end fully private ML-based data analytics systems in centralized and distributed/threshold settings. Our final objective is to implement and validate our developed schemes and prototypes through a set of use cases in machine learning. Our research results will be published in conferences, book chapters, and journals. Prototype systems as outcomes of this research will be open--sourced and can be used for knowledge translation for industry. The research program will create a unique opportunity for HQP training in an interdisciplinary research area where students will gain knowledge in cryptography, security, privacy, machine learning and IoT, which will prepare them to pursue a career in industry, academia and government. Our research will benefit sectors such as telecommunications, healthcare, banking, IT and smart electricity utilities using ML/AI.
Programme
Discovery Grants Program - Individual
Emplacement
New Brunswick, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
RGPIN-2024-04483
Name nom
Mandal, Kalikinkar
Department département
Computer Science
Organization id
75
Institution établissement
University of New Brunswick
Province en
New Brunswick
Province fr
Nouveau-Brunswick
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
31000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1507
Committee name en
Computer Science
Committee name fr
Sciences informatiques
Area of application code
802
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Information systems and technology
Area of application fr
Systèmes d'information et technologie de l'information
Research subject code
2700
Research subject group en
Information technology
Research subject group fr
Technologie de l'information
Research subject en
Information technology
Research subject fr
Technologie de l'information
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
New Efficient Cryptographic Methods and Protocols for Privacy-preserving Machine Learning
Keywords
privacy--preserving computation; cryptographic protocols; privacy-preserving machine learning; federated/collaborative learning; secure multiparty computation; lightweight cryptography; security and privacy in iot; cloud security and privacy; hybrid homomorphic encryption; trusted computing
Application summary
Cybersecurity and privacy is a growing concern in our society that includes business, people and governments. In the present age of the Fourth Industrial Revolution (Industry 4.0), extracting knowledge or value from enormous amounts of data is used to make smart decisions in business intelligence. Machine learning (ML) can learn deeper knowledge/insight from data, and in particular, deep learning has the huge potential to learn from large datasets. As more and more data makes ML powerful, unlocking values from data using ML requires sharing, processing and analyzing individual's (personal) data between organizations, which puts the individuals' rights to privacy at risk. In recent years, industries have been developing ML systems (e.g., MLaaS) for various tasks such as image classification, predictive analysis, and language models (e.g., ChatGPT). To build trustworthy and strongly secure ML systems and accelerate their deployment, security and privacy issues must be addressed, while considering various threats from adversarial machine learning attacks and cyberattacks to the ML systems. Regulations such as HIPPA and GDPR have been in place to control sharing sensitive user information, and organizations are coerced to embrace privacy enhancing technologies to protect data and users' privacy. In this research program, I propose to develop new, efficient, provably secure cryptographic techniques and security protocols to preserve the confidentiality and integrity in ML computations and develop diverse countermeasures against different cybersecurity and specialized adversarial ML attacks in an ML system. Our first objective is to design communication-efficient and secure ML training and verifiable cryptographic inference protocols to provide data and model privacy for classification and regressor models such as linear, neural networks and language models. Our second objective is to investigate new constructions of efficient hybrid homomorphic encryption schemes and multi-party threshold cryptography schemes for lightweight ciphers as well as standardized symmetric-key ciphers and develop new end-to-end fully private ML-based data analytics systems in centralized and distributed/threshold settings. Our final objective is to implement and validate our developed schemes and prototypes through a set of use cases in machine learning. Our research results will be published in conferences, book chapters, and journals. Prototype systems as outcomes of this research will be open--sourced and can be used for knowledge translation for industry. The research program will create a unique opportunity for HQP training in an interdisciplinary research area where students will gain knowledge in cryptography, security, privacy, machine learning and IoT, which will prepare them to pursue a career in industry, academia and government. Our research will benefit sectors such as telecommunications, healthcare, banking, IT and smart electricity utilities using ML/AI.